Agreement between Observed Rainfall Trends and Climate Change Simulations in the Southwest of Europe

Abstract

The lowest spatial scale at which current climate models are considered to be skillful is on the order of 1000 km because of resolution and computer capabilities. The estimation of the regional changes caused by anthropogenic emissions of greenhouse gases and aerosols therefore is problematic. Here a statistical downscaling scheme is used to study the relationship between large-scale sea level pressure and regional precipitation in southwestern Europe, both in observed data and in outputs from a general circulation model (GCM) forced with increasing levels of greenhouse gases and sulfate aerosols. The results indicate that the GCM does reproduce the main aspects of the large-to local-scale coupled variability. Furthermore, these large-to local-scale relationships remain stable in the scenario simulations. The GCM runs predict increases of advection of oceanic air masses to the Iberian Peninsula that will produce a slight decrease of precipitation amounts in the north coast and the opposite effect in the rest of the territory, with values that could reach 10 mm decade Ϫ1 in the south. In the homogenized historical records, the obtained pattern of change is very similar. These results support estimations of future regional trends simulated by the GCM under future emission scenarios.

Links between circulation and changes in the characteristics of Iberian rainfall

“…-Spain (Rodó et al, 1997) • Construction of rainfall scenarios (incorporating future changes in atmospheric composition) using a circulation-based approach to statistical downscaling: -Portugal (Corte-Real et al, 1995 -Spain (Cubasch et al, 1996;Goodess and Palutikof, 1998;Zorita and von Storch, 1999;Goodess, 2000;Palutikof et al, 2002) -Iberian Peninsula (von Storch et al, 1993Gonzalez-Rouco et al, 2000;Trigo and Palutikof, 2001). …”

Simulation of seasonal precipitation and raindays over Greece: a statistical downscaling technique based on artificial neural networks (ANNs)

“…Wilby and Wigley (2000) presented a comparative study on the predictors used for downscaling of precipitation -there is no rule for that selection as it depends on the decision of each researcher. The circulationbased predictors are the most commonly used, such as geopotential fields at different levels (500 hPa, 700 hPa), 1000-500 hPa thickness field, the sea level pressure (SLP) and specific humidity near the surface (SH) (Cavazos, 1997;Crane and Hewitson, 1998;Busuioc et al, 1999;González-Rouco et al, 2000;Trigo and Palutikof, 2001;Tatli et al, 2004).…”

“…Many different statistical methods have been employed, such as: canonical correlation analysis (CCA) (von Storch et al, 1993;Gyalistras et al, 1994;González-Rouco et al, 2000;Busuioc et al, 2001), linear and multiple regression (Kidson and Thompson, 1998;Kilsby et al, 1998;Kyselý, 2002), nonparametric models (Corte-Real et al, 1995), fuzzy logic neural networks (Cavazos, 1997;Crane and Hewitson, 1998;Trigo and Palutikof, 2001;Marzban, 2003;Knutti et al, 2003;Tatli et al, 2004), and singular value decomposition (Busuioc et al, 1999;von Storch and Zwiers, 2001).…”

Recent climatic trends in Castilla and León (Spain) and its possible influence on the potential natural vegetation

“…Numerous studies of climatic variation for the Iberian Peninsula over approximately the last two decades may be mentioned, including Hulme & Sheard (1999), González Rouco et al (2000, Sáenz et al (2001), Serra et al (2001), Rodríguez-Puebla et al (2002), Sánchez et al (2004), del Río et al (2005a), Vicente-Serrano & Moreno (2006), Mossman (2002), Mossman et al (2004), Martin et al (2004), Valero et al (2004), Múñoz-Díaz & Rodrigo (2006); del Río et al (2007).…”